We agree that whatever definition is used for surveillance, cases should be monitored to make sure they meet the case definition. Consistent definitions are necessary to detect real changes in disease rates. One of the reasons we suggest changing the surveillance definitions for staging latent syphilis is that it is very difficult to monitor all the criteria that could lead to defining a case as early latent (past syphilis test results, history of compatible symptoms, partner to an independently confirmed early syphilis case, or no other possible exposure). Monitoring titers is much easier and less subjective, as illustrated by the letter from Kohn and colleagues. They present titer analyses from all 783 early latent cases in San Francisco since 1999, but have apparently documented the criteria that lead to staging as early latent for only 293 (37%). Moreover, the information needed for staging is often not available. Persons who are tested frequently, are locatable, agree to be interviewed, and know all of their partners can supply most of the information needed for staging. Persons without sufficient information are, by default, called late latent (or perhaps unknown duration). This inherently misclassifies a fraction of cases that should have been called early latent, and that fraction will change based on the characteristics of the persons getting syphilis. Adding criteria such as response to therapy (which would again be sporadically available and operationally challenging) only increases the inherent variability of the surveillance system. Thus, no matter how well case reports are monitored, a varying percentage of infections that were recently acquired will be called late latent (or “unknown duration”) as a result of lack of information. Regardless of the reason for adding a stage called “unknown duration” to track cases with a certain age and titer, our study1 showed that many people believe that “unknown” means “not known.” Although one could argue that this is a training issue, we believe the problem is best addressed by eliminating this classification. Even if this information were correctly collected, it would not help much anyway. The main reason for surveillance is to track trends in incidence of infection. Primary and secondary cases are the best indicators of recent infections, but many jurisdictions also track trends in early latent as an indication of recent infections. Although high-titer cases do not always translate into early cases, we believe trends in high-titer cases would be less influenced by varying rates of misclassification and thus would be more helpful to monitor for trends in recent infections. Syphilis surveillance is not meant to track treatment by clinicians or intervention by the health department. There are many instances in which clinicians, disease investigators, and the staff working in surveillance should disagree about how to deal with a patient when staging is uncertain. When clinicians are uncertain about the duration of infection, they should be conservative and provide the longer treatment regimen recommended for late latent syphilis. When disease intervention specialists are uncertain, they should be conservative, assume the infection is early latent, and treat partners. Surveillance staff members should not try to force clinicians and disease interventionists to adopt the same perspective. Surveillance should endeavor to measure trends in the population as precisely as possible. We believe titer-based surveillance for latent syphilis would help.